给出了一种改进的FIR滤波器设计方法 ,该方法以复Chebyshev准则最佳地逼近一个给定的期望复值频率响应。它是通过求解半无限线性规划的对偶问题来实现的。该方法具有鲁棒性和很好的收敛特性 ,并且既可以设计复的频率响应 ,又可以设计实的频率响应。最后给出的计算机仿真结果说明了该算法的良好性能 An improved FIR filter design method is presented, which uses the complex Chebyshev criterion to optimally approximate a given expected complex frequency response. It is achieved by solving the dual problem of semi-infinite linear programming. The proposed method has robustness and good convergence characteristics, and can design both complex frequency response and real-time frequency response. Finally, the computer simulation results show the good performance of the algorithm